Data-independent acquisition proteomic analysis of the brain microvasculature in Alzheimer's disease identifies major pathways of dysfunction and upregulation of cytoprotective responses

Abstract

Brain microvascular dysfunction is an important feature of Alzheimer's disease (AD). To better understand the brain microvascular molecular signatures of AD, we processed and analyzed isolated human brain microvessels by data-independent acquisition liquid chromatography with tandem mass spectrometry (DIA LC-MS/MS) to generate a quantitative dataset at the peptide and protein level. Brain microvessels were isolated from parietal cortex grey matter using protocols that preserve viability for downstream functional studies. Our cohort included 23 subjects with clinical and neuropathologic concordance for Alzheimer's disease, and 21 age-matched controls. In our analysis, we identified 168 proteins whose abundance was significantly increased, and no proteins that were significantly decreased in AD. The most highly increased proteins included amyloid beta, tau, midkine, SPARC related modular calcium binding 1 (SMOC1), and fatty acid binding protein 7 (FABP7). Additionally, Gene Ontology (GO) enrichment analysis identified the enrichment of increased proteins involved in cellular detoxification and antioxidative responses. A systematic evaluation of protein functions using the UniProt database identified groupings into common functional themes including the regulation of cellular proliferation, cellular differentiation and survival, inflammation, extracellular matrix, cell stress responses, metabolism, coagulation and heme breakdown, protein degradation, cytoskeleton, subcellular trafficking, cell motility, and cell signaling. This suggests that AD brain microvessels exist in a stressed state of increased energy demand, and mount a compensatory response to ongoing oxidative and cellular damage that is associated with AD. We also used public RNAseq databases to identify cell-type enriched genes that were detected at the protein level and found no changes in abundance of these proteins between control and AD groups, indicating that changes in cellular composition of the isolated microvessels were minimal between AD and no-AD groups. Using public data, we additionally found that under half of the proteins that were significantly increased in AD microvessels had concordant changes in brain microvascular mRNA, implying substantial discordance between gene and protein levels. Together, our results offer novel insights into the molecular underpinnings of brain microvascular dysfunction in AD.

Keywords: Alzheimer’s disease; Blood–brain barrier; Brain microvessels; Neurovascular unit; Proteomics.

Fig. 1

Volcano plot of protein changes in AD. Positive Log2 fold-changes indicate higher levels in AD vs. control, and negative Log2 fold-changes indicate lower levels in AD vs. control. Proteins that were significantly increased were identified by the Benjamini–Hochberg method using a 5% false discovery rate. P-value significance cutoffs for each group are indicated by the horizontal red dotted line. Vertical red dotted lines indicate the smallest fold-change among significantly increased proteins

Fig. 2

Mapping of tau peptides found to be increased in AD brain microvessels. Four peptides are mapped (one mapped peptide represents two nearly identical peptides that differed by a single lysine residue at the C-terminus, see Table S2) used to compute tau protein increases in AD. Peptides (blue) were mapped to the 2N4R (441aa) Tau protein sequence and coverage includes R1-R4 sequences. Figure 2 was prepared using SnapGene

Fig. 3

Linear regression analysis of the fold-changes of significantly increased proteins in AD brain microvessels, assessed in the male and female groups. The solid line is the best-fit line of the data. The dotted line is the line of identity (slope = 1)

Fig. 4

GO analysis of significantly increased proteins in AD. A) GO Biological Process, B) GO Molecular Function. The colors of the bars reflect the -log10(FDR), with ranges shown on the heat maps to the right of each graph. Figure 4 was prepared with the ShinyGO app and with BioRender.com

Fig. 5

Functional categories analysis of the 168 significantly increased proteins in AD. Protein functions were systematically determined by checking the “Functions” section of the UniProt database, which includes references in support of protein functions. Common functional themes were then recorded, and proteins were categorized into groups and subgroups. Some proteins fell into multiple subgroups. Figure 5 was prepared with BioRender.com

Fig. 6

Cell-type enriched genes that were detected at the protein level in our dataset. Ast = astrocyte, EC = endothelial cell, PC = pericyte, MC = mural cell (genes enriched in both pericytes and smooth muscle cells), SMC = smooth muscle cell, MG = microglia, M = microglia/macrophage (genes found in both microglia and perivascular macrophages), PM = perivascular macrophage, OD = oligodendrocyte, FB = fibroblast